Since a sample is used to infer the population, the sample must be a **representative sample**. Read on to see what we mean by that!

In other words, it can’t be biased. Instead, it must represent the entire population as accurately as possible.

Among other things, statisticians could use the following guidelines to ensure that the sample is a representative sample.

First, they can make sure that the survey is not a survey **self-selected survey**.

In a self-selected survey, the participants volunteer to answer the questions.

When people volunteer, they usually have a strong opinion on the matter.

For example, suppose Candidates A, B, and C are applying for the position of superintendent in a school district.

A sample of 500 people is selected to answer questions about the candidates.

What if 100 of these participants are closely related to Candidate B and volunteer to take the survey?

It is likely that the poll is biased and does not represent the population correctly.

In addition, the **sample** can not be **too small**. Imagine that you select a sample of 200 participants from 10,000,000 residents.

It is likely that this sample does not represent the population correctly.

The sample not only has to be big enough, it also has to be **Response rate** can not be **too low**.

For example, a sample of 5000 people was selected from a population of 100,000.

The sample size looks really good. However, if only 30 people took the survey, the response rate is too low.

Sometimes the wrong people are chosen to represent the population. There are many ways this can happen.

- Choose
**just**New York State to find the median income of all Americans.

- Using
**just**the height of professional basketball players to estimate the average height of all men.

- ask
**just**Mothers who don’t believe in breastfeeding when breastfeeding is important.

Regardless of whether the sample is representative of the population or not, different studies can still produce different results.